Case study · Longitude Prize on ALS
How Oxford PharmaGenesis and King's College London are using Claude to advance research and therapeutic discovery for ALS
Executive summary
Data scientists at Oxford PharmaGenesis, the market-leading global HealthScience communications consultancy, now part of the Klick Health family, have been utilising Anthropic's Claude platform to catalyse scientific research and therapeutic target discovery in an academic collaboration with King's College London and other distinguished institutions.
Together, they are combining pioneering research, advanced bioinformatics and next-generation knowledge systems to identify therapeutic targets for amyotrophic lateral sclerosis (ALS), a relentlessly progressive and fatal neurodegenerative disease. With Claude, the team has accelerated rigorous, transparent biomedical discovery in a real-world, high-need disease area, and has laid important groundwork for next steps, including building a knowledge graph that integrates ALS-related scientific knowledge.
With Claude, we're building the first integrated sex-specific ALS knowledge graph, bringing together genetics, biology, clinical data, published research, and drug information into a single, connected framework to support precise, biologically grounded therapeutic development.
ALS is not a single biological entity, and treating it as one may obscure important signals that could lead us to new therapies. By combining large-scale genetic and epigenetic data with AI, we can begin to understand the biological differences that shape ALS and connect those discoveries to potential therapeutic targets.
Our ambition is to build a knowledge framework that does more than organise what we already know, it should help us uncover relationships we have not been able to see before, prioritise the most promising targets, and ultimately accelerate the development of more precise treatments for people living with ALS.
The challenge
ALS is the most common form of motor neurone disease (MND) with an average life expectancy of 2 to 5 years from diagnosis. The disease damages motor neurons in the brain and spinal cord, disrupting the signals that control skeletal muscles. This causes progressive muscle weakness, stiffness and wasting, affecting the person's ability to walk and speak, and eventually vital functions, such as swallowing and breathing.
There is currently no cure for ALS. The few existing disease-modifying treatments offer limited benefits and are often suitable only for small genetic subgroups. Past studies involving King's College London have proven there are differences in how ALS presents and progresses, according to biological sex. They have produced large-scale, multi-modal datasets of sex-specific data: incidence is higher in men, while women experience earlier-stage symptoms affecting speech and swallowing, as well as other qualities of disease progression and presentation.
These observations suggest there may be better approaches to identifying targets and treatments for the disease, and as disease presentation differs based upon biological sex, there could be differences in underlying disease mechanisms.
Therefore, previous study approaches that do not consider these differences will likely have diluted important signals that could reveal novel targets, more precise treatment groups, and thereby more effective and targeted treatments.
The solution
The team is developing novel knowledge-based and integrative techniques to investigate therapeutic targets that are informed by and build on the sex-specific evidence produced by an international consortium known as Project MinE. They are using Claude to accelerate the extraction, synthesis and structuring of fragmented molecular evidence, and to map relationships for integration into a large ALS knowledge graph. The team's knowledge-based and integrative methods then use this structured evidence to investigate therapeutic targets informed by the sex-specific findings from Project MinE.
Scientific methodology
Building on findings from sex-specific genetic (GWAS) and epigenetic association (EWAS) studies involving King's College London researchers and international ALS consortia, Oxford PharmaGenesis data scientists used Claude Code to support development of a reproducible bioinformatics pipeline to investigate priority genes.
The aim was not simply to assess each gene in isolation, but to structure the evidence so that it could populate a larger ALS knowledge graph: a connected, queryable map of genes, proteins, pathways and other biological evidence. For each gene, the team integrated disease association and tractability data (Open Targets), protein–protein networks (STRING), pathway context (Reactome), structural and binding-pocket evidence (UniProt, InterPro, AlphaFold and FoldSeek), tissue and cell expression (Human Protein Atlas and GTEx), and genetic regulation through cis-eQTL and sQTL data.
Claude Sonnet 4 supported the pipeline's coding and orchestration, helping the team to integrate these diverse sources and iterate the analyses rapidly. Claude Opus 4 generates structured narratives covering each gene's relevance to ALS, potential disease mechanisms, sex biology, tractability, and remaining evidence gaps. An interactive dashboard presents the resulting evidence for expert review and prioritisation.
These structured outputs are now being incorporated into a growing, queryable knowledge graph connecting genes, proteins, pathways and other biological evidence. As the graph develops, it is intended to support in silico target validation, patient-subgroup analysis, druggability assessment and explainable drug-repurposing hypotheses. Claude Opus 5 will support the next phase by identifying key biological entities and relationships and representing them as clear, interpretable facts within the graph.
Results
By analysing data separately for men and women, the underlying studies identified novel genetic risk loci and sex-associated DNA-methylation patterns in ALS. Across approximately 134,000 participants, the genetic study identified six novel autosomal risk loci that differed by sex, two in women and four in men, plus an additional ALS-associated locus on the X chromosome.
A separate DNA-methylation study of 9,274 participants identified 226 ALS-associated sites that also differed by sex, mapping to 159 genes. Seven remained significant under the more conservative Bonferroni threshold. From this broader set of findings, the current project selected six genes for systematic functional follow-up.
The Oxford PharmaGenesis pipeline and dashboard place these six priority genes within a consistent, multi-layer evidence framework, supporting expert review and prioritisation ahead of incorporation in the knowledge graph or experimental validation. Together, these tools connect population-level genetic and epigenetic findings with mechanistic and therapeutic follow-up.
The pipeline and dashboard bring together multiple layers of evidence for each candidate gene, helping researchers decide which might be most important. These outputs are also feeding into the developing knowledge graph, connecting large-scale genetic findings with potential disease mechanisms and therapeutic opportunities.
The Project MinE International Consortium team, led by King's College London in partnership with Oxford PharmaGenesis, University Medical Center Utrecht, the University of Exeter and the DEMON Network, was selected as 1 of 20 Stage 1 Discovery Award recipients in the prestigious Longitude Prize on ALS, which is principally funded by the MND Association and designed and delivered by Challenge Works, supported by Nesta and additional funders1. Each team that has progressed to Stage 1 of the Prize has been awarded £100000, as well as access to the Prize datasets and a technical support package to drive their computational work to identify and validate promising therapeutic targets.
- Footnote 1Additional funders include Nesta, the Alan Davidson Foundation, My Name’5 Doddie Foundation, LifeArc, FightMND, The 10,000 Brains Project, Answer ALS and The Packard Center at Johns Hopkins University.
Unlocking the future
Looking ahead, Claude Science will provide the workbench for building and interrogating the knowledge graph and for developing and testing hyperdimensional-computing approaches, with Claude Opus 5 supporting the underlying analytical and agentic workflows. Claude will help extract biological entities and relationships, integrate emerging Longitude Prize data, and translate outputs into interpretable insights for expert review. Together, the knowledge graph and hyperdimensional computing could reveal connections between patient characteristics, disease mechanisms and potential therapeutic targets that are difficult to identify through conventional analyses.
By combining research that accounts for biological sex with AI-assisted evidence synthesis and knowledge-graph-ready bioinformatics, the team is creating a scalable path from consortium data to prioritised, explainable ALS targets. This approach supports the team's Longitude Prize objectives and could be adapted to other progressive, life-limiting diseases.
About Oxford PharmaGenesis
Oxford PharmaGenesis is a leading HealthScience communications consultancy, founded in Oxford in 1998. Now part of the Klick Health family, the agency has more than 500 medical, scientific, and communications experts working across the UK, Europe, North America and Asia-Pacific.
About King's College London
King's College London is a world-renowned university that delivers exceptional education and world-leading research. King's College London is home to the King's MND Care and Research Centre, the first specialist centre for motor neurone disease (MND)/ALS established in the UK. The Centre has a long-standing international reputation for ALS research and has contributed to major advances in understanding the disease and developing treatments, including research that helped establish riluzole as a disease-modifying treatment for ALS. Its research spans genomics, epidemiology, biomarkers and therapeutic development, with a strong emphasis on international collaboration and data sharing.
Supplemental information and context
